More people have probably heard of Miranda v. Arizona, 384 U.S. 436 (1966), than any other criminal procedure decision. As a symbol of American justice and fair play, Miranda endures. But as a meaningful safeguard for suspects’ Fifth Amendment privilege against compelled self-incrimination, Miranda is dead. I contend that Berghuis v. Thompkins, 130 S. Ct. 2250 (2010), has fully undermined Miranda’s safeguards and will significantly alter police practices The new decision also reveals a Court that rejects Miranda’s most basic premise—that the process of in-custody interrogation contains “inherently compelling pressures” that jeopardize a suspect’s ability to exercise the privilege, Miranda, 384 U.S. at 436, 467, 478—even while the Court claims to preserve Miranda’s protections.

I. Thompkins

Let’s start with the facts. Van Chester Thompkins was arrested for murder. Two police officers sought to question him. They notified Thompkins of his rights by reading from a "Notification of Constitutional Rights " form that he refused to sign. Thompkins, 130 S. Ct. at 2256. The officers never expressly asked Thompkins if he was willing to speak with them, but simply launched into questioning. The interrogation was right out of the police manuals. It was a one-sided, accusatory monologue by officers who tried different themes to get Thompkins to talk. Id. at 2267 (Sotomayor, J., dissenting). The record of the interrogation was sparse—the officers did not use available recording equipment—though it is undisputed that Thompkins remained essentially silent. Joint Appendix at 160a, Thompkins, 130 S. Ct. 2250 (No. 08-1470), 2009 WL 4716053, at *160a. Neither officer could later remember much of what, if anything, Thompkins said during the first few hours of interrogation. However, the officers did recall that after about two hours and forty-five minutes of fruitless hectoring, Thompkins gave an incriminating response when asked if he prayed to God to forgive him for shooting the victim. Thompkins, 130 S. Ct. at 2257. The trial court admitted Thompkins’s response and the state courts affirmed his conviction. Thompkins then filed a federal habeas corpus petition, which eventually reached the Supreme Court.

The justices could have resolved the case on the limited question of whether the state courts had unreasonably applied clearly established federal law, the deferential standard used in federal habeas cases. 28 U.S.C. § 2254(d) (2006). That was the main argument of Michigan’s lawyer. Solicitor General Elena Kagan, however, filed an amicus curiae brief urging the justices to go further. Brief for the United States as Amicus Curiae Supporting Petitioner at 7-8, Berghuis v. Thompkins, 130 S. Ct. 2250 (2010) (No. 08-1470). In a 5-4 ruling, the majority followed her lead and rewrote the rules governing invocations and waivers.

II. Invocations After Thompkins

Miranda provides that to protect the Fifth Amendment privilege, warnings must be given prior to custodial questioning. If the suspect “indicates in any manner... that he wishes to remain silent” or “states that he wants an attorney,” then “the interrogation must cease.” Miranda v. Arizona, 384 U.S. 436, 473-74 (1966). Sixteen years ago, the Court ruled that a suspect who initially waives her rights but later seeks to halt questioning by asking for a lawyer must do so clearly and unambiguously. Davis v. United States, 512 U.S. 452, 459 (1994). Thompkins revises Miranda by extending Davis to the right to remain silent and even to interrogations in which a suspect did not earlier agree to talk. After Thompkins, the right to remain silent can no longer be invoked in “any manner.” Counterintuitively, a suspect must speak to claim the right to silence.

In Davis, the Court pointed to the especially powerful protections that come into play under Edwards v. Arizona, 451 U.S. 477 (1981), when someone invokes the right to counsel. See Davis, 512 U.S. at 458 (noting that Edwards held that “if a suspect requests counsel at any time during [a police] interview, he is not subject to further questioning until a lawyer has been made available”). The justices in Davis said that the strong protections of Edwards must be affirmatively invoked, and unless the defendant makes “an unambiguous or unequivocal request for counsel, the officers have no obligation to stop questioning him.” Id. at 461-62. Though the stronger Edwards protections do not apply when a suspect has only invoked the right to remain silent, the Thompkins Court found “good reason to require an accused who wants to invoke his or her right to remain silent to do so unambiguously” because this avoids difficulties of proof and provides guidance to officers. Berghuis v. Thompkins, 130 S. Ct. 2250, 2260 (2010). Nor did it matter to the Court that Thompkins—unlike Davis—had not initially agreed to speak with officers. Thus, any invocation at any stage must be articulated with precision.

Of course, a suspect still can remain mum during hours of interrogation. But to claim the right to remain silent within the meaning of Miranda—and trigger the officers’ duty to stop questioning—the suspect must affirmatively speak. That a suspect can no longer trigger the officers’ duty merely by remaining silent is plain from Thompkins, since his refusal to talk for almost three hours was insufficient to require police to curtail their questioning.

III. Waivers After Thompkins

Miranda establishes that before any statement is introduced into evidence, the government must satisfy the “heavy burden” of showing a knowing, intelligent, and voluntary waiver. Miranda v. Arizona, 384 U.S. 473, 475-76 (1966). The waiver inquiry would be the primary mechanism to assess whether the decision to speak was made free from pressure. Waiver would “not be presumed” from the fact that a confession was eventually obtained. Id. at 475. Lengthy interrogation preceding a statement would be “strong evidence” of an invalid waiver and that “the compelling influence of the interrogation finally forced” the suspect to speak. Id. at 476. Little is now left of these principles.

Thompkins shifts the burden of proof of waiver or, at least, makes the burden featherweight. The Court held that “[w]here the prosecution shows that a Miranda warning was given and that it was understood by the accused, an accused’s uncoerced statement establishes an implied waiver of the right to remain silent.” Thompkins, 130 S. Ct. at 226. Since the prosecution already must prove that warnings were given and that a voluntary statement was made in order to admit the statement, the only additional requirement for waiver is that the accused understood the warnings. But the Court made clear that it is primarily the defendant’s burden to show a lack of understanding. While the majority noted that Thompkins could read English and had time to read the warnings, the justices emphasized first and foremost that “there is no contention that Thompkins did not understand his rights; and from this it follows that he knew what he gave up when he spoke.” Id.

The majority found an implied Miranda waiver on an extreme set of facts. Rejecting the argument that officers must obtain a Miranda waiver before questioning a suspect, a waiver can now be obtained in the middle of an interrogation: “The fact that Thompkins made a statement about three hours after receiving a Miranda warning does not overcome the fact that he engaged in a course of conduct indicating waiver.” Id. at 2263.

The relocation of waiver into the heart of an interrogation and the finding of implied waiver after hours of questioning remake the waiver doctrine. In North Carolina v. Butler, the only prior Supreme Court implied waiver case, the defendant agreed to talk with police but balked at signing a written form. 441 U.S. 369, 371 (1979). The Court found that waiver can be inferred from actions and words, and remanded for the state courts to decide whether Butler had given up his rights. Id. at 373, 376. Since Butler, all federal circuits, as well as appellate courts in at least forty-two states and the District of Columbia, have upheld the admission of statements based upon implied Miranda waivers. Charles D. Weisselberg, Mourning Miranda, 96 Cal. L. Rev. 1519, 1581-83 (2008). But in Thompkins, neither Michigan nor the Solicitor General were able to cite any decision in which a court found that a suspect had given an implied waiver after lengthy questioning. Thompkins persevered for almost three hours before succumbing to his interrogators. In finding a waiver on these facts, Thompkins gives us an implied waiver doctrine on steroids.

IV. Ignoring Miranda’s Premise

What is missing from the Thompkins majority’s opinion is any understanding of how the process of interrogation influences suspects’ behavior. The justices wholly disregarded what happened in the hours between the warnings and Thompkins’s statement.

This disregard is contrary to Miranda’s basic premise: modern custodial interrogations contain inherently compelling pressures that jeopardize the privilege. These pressures are the very reason safeguards are needed at all. In Miranda, Justice White asked how we could assume that those same compelling pressures would not taint waivers. Miranda v. Arizona, 384 U.S. 436, 536 (1966) (White, J., dissenting). As a practical matter, the Miranda Court alleviated this concern by requiring the safeguards to be employed before questioning begins. But if waiver can be found after hours of pressure-filled interrogation, surely Justice White’s question becomes central. Yet the Thompkins majority failed even to consider how Thompkins’s lengthy interrogation led to waiver. Rather, the Court only asked whether his statement was coerced under traditional voluntariness standards, saying that “Thompkins does not claim that police threatened or injured him during the interrogation or that he was in any way fearful.” Berghuis v. Thompkins, 130 S. Ct. 2250, 2263 (2010). Further, the majority insisted, “there is no authority that an interrogation of this length is inherently coercive.” Id.

This lack of concern about the validity of a waiver given after hours of interrogation is perhaps the most disturbing aspect of Thompkins. The majority’s opinion reflects a rejection of Miranda’s core premise, or at least a willingness to discount it entirely. That leaves the Miranda doctrine at war with itself, for without faith in Miranda’s premise, there is no need for any safeguards at all.

V. Are Miranda’s Safeguards Effective?

But if we accept (as I do) the Miranda Court’s determination that custodial interrogations contain compelling pressures, does today’s Miranda doctrine afford suspects a meaningful ability to choose between speech and silence? It is difficult to conclude that it does.

To begin, there are good reasons to doubt that any set of standardized warnings or one-size-fits-all safeguards could be uniformly effective. Although the Miranda Court had police manuals to assess the impact of modern interrogation techniques, there was no empirical basis for the justices’ conclusion that a system of warnings and waivers could actually counter the “inherently compelling pressures” of custodial interrogation. Miranda’s prophylaxis “rested upon an untested, unverified, and unproven assumption...that [warnings]... work.” Morgan Cloud et al., Words Without Meaning: The Constitution, Confessions, and Mentally Retarded Suspects, 69 U. Chi. L. Rev. 495, 517 (2002). The social science literature now shows that we cannot assume that warnings effectively enable suspects to understand their rights, given the varying language of the warnings and the varying capabilities of suspects, among other reasons. Saul M. Kassin et al., Police-Induced Confessions: Risk Factors and Recommendations, 34 Law & Hum. Behav. 3, 7-9 (2009) (White Paper adopted by the American Psychology-Law Society); Weisselberg, supra, at 1563-77.

Regardless of whether Miranda’s safeguards could ever have been effective, decisions predating Thompkins significantly undermined them. As Colorado v. Spring made clear, officers do not have to tell suspects which crimes they are suspected of committing. 479 U.S. 564, 576 (1987). Nor must they tell suspects that an attorney is trying to contact them. Moran v. Burbine, 475 U.S. 412, 425 (1986). Yet surely this is information suspects would want to know before deciding whether to speak. Moreover, the warnings do not have to contain any particular language. Florida v. Powell, 130 S. Ct. 1195, 1206 (2010). Officers can sometimes tell suspects that they are free to leave and make it appear as if even a stationhouse interrogation is noncustodial, thereby manipulating the boundary of Miranda’s protections. Cf. Stansbury v. California, 511 U.S. 318, 326 (1994) (per curiam) (“[O]fficers’ subjective and undisclosed suspicions...do not bear upon the question whether [a suspect is] in custody, for purposes of Miranda, during the station house interview.”); California v. Beheler, 463 U.S. 1121 (1983) (per curiam) (holding that Miranda warnings are not required when a suspect voluntarily comes to the stationhouse, is told he is not under arrest, and is allowed to leave voluntarily).

I am not the first to suggest that Miranda’s heralded safeguards fall short. The late Welsh White (among others) made this point with particular force almost a decade ago. Welsh S. White, Miranda’s Waning Protections (2001). Richard Leo—perhaps our country’s most ardent scholar of the interrogation process—explains that “the Miranda ritual makes almost no practical difference in American police interrogation.” Richard A. Leo, Police Interrogation and American Justice 124 (2008). He goes on to say that officers “have learned how to ‘work Miranda’ to their advantage” and concludes that “once waived, Miranda does not affect the subsequent interrogation because it does not prohibit any post-waiver techniques.” Id.

Two years ago, I wrote that as a prophylactic device to protect suspects’ Fifth Amendment privilege, Miranda was effectively dead. The causes were varied, but included the then-growing use of implied waivers, and aggressive applications of Davis and Beheler—all of which allowed officers to limit the reach of Miranda and to relocate warnings and waivers to the interior of a highly-structured interrogation process. I expressed doubts that the Supreme Court could or would repair it. Weisselberg, supra, at 1590-94. Thompkins has resolved any lingering doubts more swiftly and surely than I could have foreseen.

VI. Thompkins and Police Training

Thompkins has enormous practical implications for policing. Many law enforcement agencies have already trained police about implied waivers. The "Notification of Constitutional Rights" form used by the officers in Thompkins was designed to set up an implied waiver: the form asked Thompkins to acknowledge that he understood his rights but nowhere asked him to waive them. California’s standard Miranda card sets up an implied waiver—though the back of the card contains some suggested questions for officers who opt to seek an express waiver.

Thompkins is likely to change police training in at least three respects. First, not all law enforcement agencies currently promote implied waivers. Thompkins will give officers far greater comfort in seeking implied waivers than under prior law, and more agencies will use such waivers. Second, given that a waiver was found under the extreme facts of the Thompkins interrogation, we can expect police to be told that a waiver may be implied even after hours of questioning unless the defendant can prove that the statement was involuntary under traditional standards. This message is quite different from current training. Third, officers will be told that Davis applies to all types of invocations and even at the outset of questioning. Considering how stingy courts have been in finding an unambiguous request under Davis (see, for instance, Berghuis v. Thompkins, 130 S. Ct. 2250, 2277-78 (2010) (Sotomayor, J., dissenting)), police will be trained they can give warnings and interrogate until a suspect makes a pristine invocation of her rights.

VII. Miranda Remains—Sort of

What does Miranda do today? A decade ago, in turning aside a direct attack on Miranda, Chief Justice Rehnquist wrote that Miranda’s warnings “have become part of our national culture.” Dickerson v. United States, 530 U.S. 428, 443 (2000). True enough. Miranda remains a potent symbol for the American people. But symbolism aside, as a practical matter Miranda functions largely as a safe-harbor rule for police. Officers know that if they comply with Miranda’s formal strictures, a challenge to the voluntariness of a statement is likely to fail.

It may be too difficult politically for the Court to overturn Miranda even if the majority rejects its central premise. So, instead, like physicians who prescribe a placebo because they disbelieve a patient’s earlier diagnosis, the justices have given us ineffective procedures to address a problem that they do not think exists. One could wish they had addressed the diagnosis directly, which would have opened a much larger and welcome debate about the interrogation process and the law.

Miranda may be a hollow symbol, as Professor Weisselberg mourns, but that is nothing to lament. If Thompkins is unfaithful to Miranda, that is only because Miranda was unfaithful to the Fifth Amendment’s requirement of compulsion. Miranda’s protections map poorly onto the kinds of compulsion that produce false confessions and the categories of people likely to confess falsely. Miranda thus shielded some savvy, guilty recidivists while doing little to protect juveniles, the mentally retarded, and other innocent defendants most likely to confess. Thompkins is just another step toward jettisoning the failed Miranda experiment. The hard question now is how to adopt better safeguards tailored to what we now know about false confessions.

I. Invocations

Professor Weisselberg first laments that “a suspect must speak to claim the right to silence.” That claim is specious. All he need do is simply keep his mouth shut. He also has the option, if the questioning grows too heated, of invoking his rights midstream. As Bill Stuntz has argued, Miranda is really a way of giving defendants a tool they can invoke if questioning gets to be too overbearing. William J. Stuntz, Miranda’s Mistake, 99 Mich. L. Rev. 975, 981-88, 991 (2001). Empirical evidence suggests that very few suspects do that; they either invoke their right to silence immediately or else speak. Id. at 988. But we don’t really know what that result means. The mere presence of waivers or absence of invocations doesn’t tell us that Miranda is doing nothing; maybe police are just staying away from tactics that push warned suspects into invoking their rights. It could be that police have long since abandoned the third-degree tactics that were infamous early in the last century, deterred in part by Miranda, so there is much less need for a shield against true coercion. It could also be that many defendants don’t understand their rights well enough to invoke them midstream. We don’t know how important each of these factors is, though studies do show that adolescent and mentally disabled suspects don’t understand their rights well. What we do know, however, is that recidivists are far more likely to invoke Miranda rights in the first place. Id. at 993. Miranda thus gives those most likely to be guilty a windfall while doing little to shield the innocent neophyte from compulsion.

The failure of defendants to invoke their rights midstream is troubling only if we assume with the Miranda Court that stationhouse interrogations inherently or presumptively involve compulsion within the meaning of the Fifth Amendment. But there is little reason to think so. There is little evidence that police use fists, rubber hoses, or similar physical force in interrogation except in the rarest of cases; if they did, any confessions would be blatantly involuntary and inadmissible. Nor is there much evidence of using concrete threats or promises, as opposed to vague words of soft hope. There is room to question prolonged interrogations involving deprivation of food and sleep, but Thompkins expressly noted that it was not addressing such a case. Berghuis v. Thompkins, 130 S. Ct. 2250, 2254 (2010).

II. Waivers

Nor am I troubled that Thompkins diluted Miranda’s strong presumption against waiver. Waiver is simply a prophylaxis, creating additional evidence that a defendant understood her rights and chose not to exercise them. A defendant who does not waive may not understand English, or may be so intoxicated or mentally ill that any confession would be thoroughly unreliable. But particularly when an adult of average intelligence hears and understands her rights, there is little reason to presume that any statement is compelled absent an affirmative waiver.

III. Stretching the Idea of Compulsion

At the root of Miranda and of Professor Weisselberg’s objections to Thompkins is an overbroad understanding of compulsion. He shares Miranda’s assumption that custodial interrogations are inherently compelling and criticizes Thompkins for focusing narrowly on the lack of threats, injuries, fears, or prolonged interrogation.

Instead, Professor Weisselberg objects that there “is information that suspects would want to know before deciding whether to speak” and that police are “manipulating the boundary of Miranda’s protections.” These claims are unmoored from the Fifth Amendment’s requirement of compulsion, particularly the type of compulsion that could lead to false or unreliable confessions. They sound more like a sporting theory of justice: Bentham memorably described the desire to allow those who are clearly guilty an opportunity to escape blame as giving the fox in a fox hunt a fair chance of escaping the hunter. 5 Jeremy Bentham, Rationale of Judicial Evidence 238-39 (Fred B. Rothman & Co. 1995) (1827). But criminal justice is not a sport; it ought to be about seeking the truth. Convicting the guilty is an imperative, as is freeing the innocent. And absent from both Miranda and Professor Weisselberg’s critique is any differentiation of pressures that could lead to false confessions from those that generally produce true ones.

Our justice system depends in part on reliable confessions. We should not stand in the way of interrogation techniques that produce truthful confessions so long as they do not create an unacceptable risk of producing false ones. As Justice Scalia memorably put it,

it is wrong, and subtly corrosive of our criminal justice system, to regard an honest confession as a “mistake.”... Not only for society, but for the wrongdoer himself, admissio[n] of guilt... , if not coerced, [is] inherently desirable because it advances the goals of both justice and rehabilitation.... We should... rejoice at an honest confession, rather than pity the “poor fool” who has made it....

Minnick v. Mississippi, 498 U.S. 146, 167 (1990) (Scalia, J., dissenting) (citations omitted) (internal quotation marks omitted).

If that is right, then Miranda’s lack of impact on psychological interrogation methods, and police exploitation of Miranda’s loopholes, should be not mourned but celebrated. If criminal justice is a search for the truth, then Miranda should be no obstacle. Leaving Miranda as a hollow symbol, dead on the inside, may be the best stable compromise that one can make with a flawed decision. Police have learned to live with it, so in practice it costs us few confessions.

IV. Involuntary, False Confessions

Professor Weisselberg is on stronger ground complaining that Miranda warnings have become a safe harbor for police. Miranda meant for its warnings and waivers to be necessary conditions of interrogation, not sufficient ones. Interrogations were still supposed to satisfy the traditional due process standards of voluntariness. But, in practice though not in theory, lower courts have turned Miranda’s bright-line rule into a safe harbor—a sufficient condition for admission and not just a necessary one.

Empirical evidence confirms that many involuntariness factors are present in false confession cases, yet judges admit the statements. The bulk of false confessions comes from teenaged, mentally retarded, or mentally ill suspects. A recent empirical study of defendants who confessed falsely but were later exonerated found that thirty-three percent were juveniles and forty-three percent were mentally disabled, while others were clearly emotionally disturbed. Brandon L. Garrett, The Substance of False Confessions, 62 Stan. L. Rev. 1051, 1064, 1094-95 (2010). Most of the exonerees were questioned for more than a day (with breaks only for sleep and food) or multiple times over several days, sometimes for as long as thirty to forty hours over a week. Id. at 1095. Yet every one of these forty false confessions was ruled voluntary and admissible. Id. at 1054. Miranda, in essence, has become a substitute for serious voluntariness scrutiny.

Miranda’s hollowing out, if not demise, should prompt us to think hard about how to prevent these injustices that have persisted alongside Miranda for decades. We academics tend to put too much faith in judicial review and rules, but case-by-case litigation alone has proven impotent to ferret out and prevent these abuses. And instead of selling interrogation reform as a matter of fair play, reformers should focus squarely on measures to protect the innocent and prevent wrongful convictions.

V. Better Reform Ideas

A prime aim of reform ought to be full videotaping of all interrogations. Though many police departments record some interrogations, they sometimes do so selectively, turning on the tape only for the final confession or turning off the tape when a suspect gives a problematic response. Automatic, tamper-proof, voice-activated recording equipment with time and date stamps could continuously record everything that goes on in an interrogation room. That is crucial because, in most of the false confessions studied, police appear to have disclosed to the suspects (intentionally or inadvertently) nonpublic details that were crucial to making their confessions credible. Id. at 1068-74. Full videotaping cannot prevent every abuse—police could still taint confessions at the crime scene or in the patrol car, and equipment occasionally malfunctions. But videotaping would greatly reduce the opportunities for coercion and suggestion and, when they do occur, reveal them to judicial scrutiny. And though police departments initially resist adopting full videotaping, once they begin using it, most embrace recording. They learn to love recording because it protects them against false claims of coercion, intimidation, and abuse; obviates constant note taking; forecloses motions to suppress; and encourages guilty pleas. See Thomas P. Sullivan, Nw. Univ. Sch. of Law Ctr. on Wrongful Convictions, Police Experiences with Recording Custodial Interrogations 6-19 (2004), http://www.innocenceproject.org/docs/Police_Experiences_Recording_Interrogations.pdf.

As Professor Garrett suggests, other reforms to interrogation techniques are both simple and promising. Police could strive to keep many details of crimes nonpublic and note in the file which facts had been held back from the press and public. They could also use double-blind techniques, having interrogators other than the detectives on the case question suspects without knowing the nonpublic details, so any confessions would not be tainted by suggestion. These double-blind detectives could even have suspects pick victims out of photo arrays, thus excluding those without personal knowledge. Garrett, supra, at 1115-16.

Ultimately, the most important step is to do more research and craft special safeguards for teenagers, the mentally retarded, and the mentally ill or disturbed, who account for most false confessions. Though courts are unlikely to forbid long or deceptive interrogations in general, they could adopt special rules for these vulnerable classes of suspects.

VI. How to Achieve Reforms?

It is easier to spell out potential reforms than to figure out how to implement them in America. Miranda turned interrogation regulation into countermajoritarian combat, with judges trying to thwart reliable interrogation techniques and thus the public’s interest in crime control. Now, however, innocence commissions have changed the political dynamic, tapping into the admirable fear of executing an innocent man. That is a far more politically appealing and thus more durable basis on which to ground interrogation reform.

So, perhaps I can leave that issue hanging and solicit Professor Weisselberg’s response: What strategies can persuade police departments, prosecutors’ offices, and legislatures that they need and want to pursue interrogation reform? Can judges, instead of creating intricate and unproven interrogation codes themselves, somehow prompt police self-regulation and best practices? And, in particular, what measures would work best to preserve reliable interrogation of defendants who are young, mentally retarded, mentally ill, or emotionally disturbed while guarding against suggestion and exploitation? Those are far different questions from the ones posed by Miranda’s blanket hostility to interrogation, but ultimately they should prove more fruitful.

I had hoped Professor Bibas would argue that, post-Thompkins, Miranda’s procedures actually do protect the Fifth Amendment privilege against compelled self-incrimination. I would root for him to win that argument because I would like to live in a world in which a relatively simple set of procedures affords suspects a meaningful opportunity to choose between speech and silence in the stationhouse. But Professor Bibas agrees with me that we do not live in that world, if indeed we ever did. While I call Miranda’s protections “dead” and he prefers the term “hollow,” we both see the need to move on and think about other ways to regulate the process of police interrogation.

However, moving on requires some agreement about what we want to protect as well as an understanding of how modern interrogation actually works. We cannot simply leave Miranda behind, for that decision has a lot to say on both of these points. Professor Bibas appears to disagree with the Miranda Court (and me) on two fundamental aspects of the ruling: first, that the Fifth and Fourteenth Amendments protect basic values and aspects of our criminal justice system other than the need to avoid false confessions; and second, that modern interrogation techniques do contain inherently compelling pressures that endanger suspects’ rights and that may undermine some of these objectives of our system. I will address both points and then suggest possible reforms.

I. The What—Reliability and a Whole Lot More

Professor Bibas contends that Miranda was unfaithful to the Fifth Amendment’s notion of compulsion, arguing that the Court failed to differentiate between pressures leading to false confessions and pressures that generally produce true confessions. We should examine both Fourteenth Amendment coercion and Fifth Amendment compulsion (concepts that have generally been merged).

Voluntariness within the meaning of the Due Process Clause turns on “whether the behavior of the State’s law enforcement officials was such as to overbear petitioner’s will to resist and bring about confessions not freely self-determined... with complete disregard of whether or not petitioner in fact spoke the truth.” Rogers v. Richmond, 365 U.S. 534, 544 (1961). A statement by a defendant “might be proved to be quite unreliable, but [that] is governed by the evidentiary laws of the forum,... and not by the Due Process Clause.” Colorado v. Connelly, 479 U.S. 157, 167 (1986). The purpose of the voluntariness requirement is “to prevent fundamental unfairness in the use of evidence, whether true or false.” Id. (internal quotation marks omitted).

Now to Miranda and the Fifth Amendment: “[T]he basic purposes that lie behind the privilege against self-incrimination do not relate to protecting the innocent from conviction, but rather to preserving the integrity of a judicial system in which even the guilty are not to be convicted unless the prosecution ‘shoulder the entire load.’” Tehan v. United States ex rel Shott, 382 U.S. 406, 415 (1966). The Fifth Amendment furthers a “complex of values,” which include respecting the dignity and integrity of citizens, maintaining a fair balance between an individual and the state, and demanding under our adversary system that the government produce evidence by its own labors, “rather than by the cruel, simple expedient of compelling it from [a suspect’s] own mouth.” Miranda v. Arizona, 384 U.S. 436, 460 (1966) (citation omitted). While the history of the privilege and many of these values have been well debated (see, for example, Albert W. Alschuler, A Peculiar Privilege in Historical Perspective, in The Privilege Against Self-Incrimination 181 (R.H. Helmholz et al. eds., 1997)), the purpose of the privilege is not to prevent false confessions. To be sure, it is sometimes noted that a benefit of Miranda is that it helps avoid the use of unreliable statements. E.g., Withrow v. Williams, 507 U.S. 680, 692 (1993). But the Court’s Miranda decisions emphasize that the goal is to protect autonomy: “[T]he fundamental purpose of... Miranda [is] ‘to assure that the individual’s right to choose between speech and silence remains unfettered throughout the interrogation process.’” Connecticut v. Barrett, 479 U.S. 523, 528 (1987) (emphasis added) (quoting Miranda, 384 U.S. at 469). Perhaps we care about autonomy because, as Professor Peter Brooks writes, “[t]he law wants and needs a model of human agents as free and rational decision-makers, even in the confession of guilt.” Peter Brooks, Troubling Confessions 76 (2000).

Contrary to Professor Bibas’s suggestion, an effort to further the Fifth Amendment’s legitimate values and objectives ought not be demeaned as promoting some “sporting theory of justice.” Applying the Fifth Amendment in the stationhouse requires us to face difficult and serious questions. We must consider the extent to which individuals should be allowed to exercise free will in deciding whether to speak, the amount of information they should be provided before making life-altering decisions, and whether we want suspects to be the same free and rational decisionmakers that we assume them to be when we assess criminal culpability. These are not questions of sport.

I am also uncertain about the implications of Professor Bibas’s theory of coercion and compulsion. He seems to accept that a statement is involuntary (and a Miranda waiver is involuntary) if one extracts it by physical force or concrete threats or promises. These are obvious and traditional markers for situations in which a person’s will is overborne, yet the concept is broader than these examples. And if Professor Bibas also requires a showing that an officer’s action is likely to lead to a false confession, even physical abuse might not suffice. We have “strong evidence that coercive techniques increase the odds of a false confession... but we do not know by how much. It is possible, for all we know, that the overwhelming majority of coerced confessions are true.” Samuel R. Gross, Convicting the Innocent, Ann. Rev. L. & Soc. Sci. 173, 188 (2008). Requiring a causal link between abhorrent tactics and reliability would undermine our ability to condemn tactics that have no place in a civilized society.

II. The How—Modern Police Interrogation

I have argued that Thompkins implicitly rejects the idea that custodial interrogations contain inherently compelling pressures. Professor Bibas also rejects this premise, but does so expressly. I am grateful for that, because his candid views further a conversation that is essential to the process of reform.

On inherent compulsion, Miranda got it right. We now know much more about the process of interrogation, and the data support the Miranda Court’s conclusion. To begin with, officers do use the techniques in police training manuals. Saul M. Kassin et al., Police Interviewing and Interrogation: A Self-Report Study of Police Practices and Beliefs, 31 Law & Hum. Behav. 381, 384 (2007); Weisselberg, supra, at 1530-37. Although Professor Bibas would like more research on interrogation tactics and vulnerable populations, let’s begin with the research that we already have.

It is true that interrogation tactics may affect certain suspects more than others. But we also know that the techniques are themselves quite powerful, even on populations of suspects whom we do not ordinarily consider vulnerable. The techniques increase suspects’ anxiety, instill a feeling of hopelessness, and distort suspects’ perceptions of their choices by leading them to believe they can escape only by cooperating with investigators—which is essentially a “fraud.” Richard A. Leo, Police Interrogation and American Justice 25 (2008). Interrogation is a carefully constructed, guilt-presumptive process. Researchers have shown that maximization techniques convey implied threats of punishment, and minimization techniques successfully convey implied promises of leniency. Saul M. Kassin & Karlyn McNall, Police Interrogations and Confessions: Communicating Promises and Threats by Pragmatic Implication, 15 Law & Hum. Behav. 233, 234-35 (1991). The pressures of interrogation are manifest: one manual lists “Psychological Domination” as the first step of a confrontational interrogation. Weisselberg, supra, at 1534. Even if one is most concerned with false confessions, as Professor Bibas seems to be, research has identified situational risk factors associated with false confessions (such as interrogation time, presentation of false evidence, and use of common minimization techniques) as well as dispositional risk factors (such as age and mental impairment). Kassin, Police-Induced Confessions, supra at 16-22. Though there are more studies to discuss, this research tells us that the search for coercion or compulsion must involve more than a cursory look for markers such as violence and express threats and promises, and that the problem of coerced or compelled confessions is not confined to a few vulnerable populations.

I understand that the Supreme Court does not welcome this news. “Where psychology brings ambiguity and complexity and layering of motive, the Court wants certainty and bright lines.” Peter Brooks, The Future of Confession, 1 Law, Culture & the Human. 53, 60 (2005). Miranda was a failed effort to provide bright lines and avoid more taxing and nuanced determinations of voluntariness. If we are honest, however, we cannot shut our eyes to the reality that interrogations are structured as they are because they lead—and often compel—suspects to talk. Because Miranda is an empty ritual and effectively forecloses a meaningful voluntariness inquiry, I have argued that we would be better off with a full voluntariness inquiry than with Miranda. But the inquiry must consider the effect of interrogation techniques—even those Professor Bibas would call “soft.” Psychological coercion includes “police use of interrogation techniques that, cumulatively, cause a suspect to perceive that he has no choice but to comply with the interrogators’ demands.” Leo, supra, at 230. We cannot simply look for a single marker of coercion or compulsion. The concept of voluntariness is more complicated than that.

III. Reform

In addition to judicial oversight through richer and more nuanced voluntariness determinations, I favor legislative and administrative reforms—though I also struggle with how to create incentives for reform and self-regulation. One possibility is to offer a trade: Miranda’s procedures may be replaced with equally effective alternatives, see Dickerson v. United States, 530 U.S. 428, 440 (2000), a low bar to clear since Miranda’s procedures are themselves singularly ineffective. Although some officers may prefer to have a “hollow” rule that operates as a safe harbor, others in law enforcement may be willing to explore reforms if the payoff is dispensing with Miranda’s procedures altogether. At least this trade-off is worth exploring.

So what might these reforms be? I agree with Professor Bibas and many other commentators that videotaping interrogations is essential. Recording may dampen extreme tactics. In New South Wales, recording entirely ended the practice of “verballing,” which consisted of officers fabricating confessions. David Dixon, Interrogating Images 1, 262 (2007). While videotaping creates a factual record, it is not a panacea. Some images may prejudice the defendant. We must also ensure that all of the interrogation is recorded, and that cameras show both the officer and the suspect. Id. at 3, 263-75. Camera angles affect the factfinder’s assessment of a statement. Daniel Lassiter et al., Videotaped Interrogations and Confessions: A Simple Change in Camera Perspective Alters Verdicts in Simulated Trials, 87 J. Applied Psychol. 867, 867-69 (2002). Moreover, we still need to assess the legitimacy of what is recorded in each case by some acceptable standard.

I would substantially restrict false evidence ploys and deception about evidence. In addition to documented concerns about false confessions, it is important to build trust between officers and members of the community. The state does not charge all suspects in police investigations, and deception about evidence undermines trust. It may also have a corrosive effect on law enforcement.

I would also restrict the use of minimization techniques, which often contain implied promises of leniency. Not only are they a risk for false confessions, they may lead to a false postadmission narrative, even when the defendant is guilty of the offense. Officers are trained to elicit a confession, a suspect’s story, after obtaining an initial admission. Yet officers often get that breakthrough admission by suggesting minimizing themes without any regard to their truth. Police are trained to develop different themes, and that what is most important is that the themes be acceptable to the suspect. One training video advises officers to prepare at least five potential themes before an interrogation. DVD: Interrogation Techniques Telecourse (Cal. Comm’n on Peace Officers Standards and Training 2005) (on file with the University of Pennsylvania Law Review); see also Fred E. Inbau et al., Criminal Interrogation and Confessions 239 (4th ed. 2001) (“When switching to a different theme, the investigator should not indicate disappointment for having presented the first theme. He should just quickly embark upon another.”). While I am concerned with minimization tactics themselves, I am also troubled because officers develop themes solely to appeal to suspects. A defendant who has accepted an officer’s theme may minimize the wrongfulness of her conduct by falsely implicating others.

Nor are the harms limited to the possibility of wrongfully convicting other people. I can only begin to imagine the pain experienced by a rape victim and her family who watch the videotaped interrogation of a defendant who admits the sexual act but accepts the detective’s suggestion that the victim “acted like she might have been a prostitute;” or that “[i]f she hadn’t gone around dressed like that [the defendant] wouldn’t be in this room now;” or that the ten-year-old victim was “well advanced for her age.” Inbau, supra, at 256-57.

Other reforms may be useful, but these are a start. I am very glad for this dialogue, and am grateful to Professor Bibas and the PENNumbra editors. I look forward to Professor Bibas’s closing contribution.

Professor Weisselberg vigorously defends Miranda’s expansive reading of the Fifth Amendment’s purposes as well as Miranda’s strong presumption of compulsion in police interrogation. The privilege, he argues, is much more about protecting guilty suspects’ autonomy and dignity than about protecting the innocent. And the power of modern interrogation techniques, he claims, shows that they compel suspects just as torture, threats, and promises do.

While his claims are faithful to Miranda and its progeny, they are unmoored from the text and history of the privilege against compelled self-incrimination. Moreover, they are normatively unattractive, reaching far beyond unreliable statements and torture to protect savvy recidivists from vigorous inquiry into their guilt. Finally, Miranda’s overbreadth has distracted attention from reforms targeted at vulnerable suspects and unreliable tactics.

On a positive note, Professor Weisselberg and I both come to bury Miranda, not to praise it. He is under no illusions that Miranda ever worked, let alone that it can be propped up now. He offers some useful suggestions to reform, suggestions on which I will try to build.

I. The Purposes of the Privilege Against
Compelled Self-Incrimination

First, Professor Weisselberg attempts to bolster Miranda’s reading of the Fifth Amendment by quoting a number of its progeny. To borrow from Wittgenstein, that is like trying to verify a newspaper by checking it against another copy of itself. We need to go back to the history of the Fifth Amendment and the abuses against which it was designed to guard. Too often people use “the privilege against self-incrimination” as shorthand, omitting the key word “compelled” from its limits. Thus, Professor Weisselberg embraces a gauzy web of soft values such as judicial integrity, individual dignity and autonomy, privacy, and making the government shoulder the whole burden of proving guilt in an adversarial system. He is in good company—the Warren Court and many commentators have likewise overread the privilege.

No serious historian believes that the privilege was adopted for such a hash of individualist, modern reasons. The English, as well as American colonists, feared confessions extracted by torture or the compulsion of having to swear an oath, as the Spanish Inquisition and the Star Chamber had done. They probably would have put threats of future punishment or promises of leniency into those categories as well. Alschuler, supra, at 192. But they never would have thought, as the Warren Court did, that simply drawing an adverse inference from a defendant’s silence is a form of compulsion tantamount to torture or threats. See, e.g. Griffin v. California, 380 U.S. 609, 614-25 (1965). On the contrary, the privilege allowed judges and magistrates to interrogate suspects forcefully, so long as they were not under oath. Alschuler, supra, at 193.

In England and colonial America, justices of the peace routinely interrogated suspects out of court and drew adverse inferences from their silence. Eben Moglen, The Privilege in British North America: The Colonial Period to the Fifth Amendment, in The Privilege Against Self-Incrimination, supra, at 109, 117-22. Professor Weisselberg would presumably criticize investigators’ drawing of adverse inferences as an implied threat. But the suspects were not forced to be witnesses within the meaning of the Fifth Amendment. They did not have to testify at trial on pain of torture or comparable compulsion, though any pretrial confession was admissible at trial. Id.

So understood, the privilege against compelled self-incrimination was consistent with the colonists’priorities. As Akhil Amar has convincingly argued, following Alexis de Tocqueville, the Bill of Rights was not some countermajoritarian imposition of unpopular individual rights, but was “fundamentally populist and majoritarian.” Akhil Reed Amar, The Bill of Rights: Creation and Reconstruction 88 (2000). It was not designed to shield guilty suspects from deserved punishment, unless authorities used methods tantamount to torture.

The gauzy tangle of rationales upon which Professor Weisselberg relies is unappealing. As my Rebuttal noted, Miranda’s overbroad prophylaxis winds up sheltering few innocent or neophyte defendants, but mostly savvy recidivists who are far more likely to invoke their Miranda rights. Moreover, Professor Weisselberg’s and Miranda’s complex of rationales is directly at odds with other foundational principles of Fifth Amendment law. Lawyers for the government can force parents to testify against their children, can subpoena witnesses to reveal embarrassing truths, and can force them to disclose the deepest, most incriminating secrets possible so long as they provide immunity. Prosecutors can issue subpoenas or use search warrants to confiscate incriminating documents and tangible objects. And they can even invade a defendant’s bodily integrity, extracting blood or DNA to prove the defendant’s own guilt. See Schmerber v. California, 384 U.S. 757, 765 (1966) (holding that the Fifth Amendment privilege does not forbid forcibly extracting a defendant’s blood and admitting the resulting analysis into evidence). In other words, the government can use defendants in all other kinds of ways to help prove their own guilt and invade their privacy and autonomy. The real distinction is that blood samples, fingerprints, subpoenaed documents, evidence from searches, and subpoenaed testimony are reliable evidence, while coerced confessions are not. Thompkins, by cutting back Miranda’s reach, brings the regulation of interrogations more into line with these other areas of Fifth Amendment case law.

II. Interrogation Is Seduction, Not Compulsion

The ban on torture is not limited to reducing unreliable statements, though that may be one of its key purposes. But Professor Weisselberg cannot plausibly claim that the pressures of stationhouse interrogation are remotely analogous to torture. Suspects may speak, and they may regret speaking. But today, interrogation is far more a matter of seduction than compulsion. The distinction is an important one. Seduction operates through a defendant’s free, autonomous choice, to use Professor Weisselberg’s values. It may establish a false rapport; it may play on emotions, sympathies, guilt, and remorse. But those kinds of social influences are such a far cry from the Wickersham Commission’s third degree that one cannot lump them with physical violence and concrete threats and promises. They involve lawful pressures, not force or fraud. I am unpersuaded by Professor Weisselberg’s suggestion that most interrogation involves implied threats tantamount to force or implied promises rising to the level of fraud. There is little reason to expand the privilege’s truth-defeating function—to paternalistically protect guilty suspects from revealing their guilt in ways that they may regret but we should celebrate.

III. Remaining Causes for Concern

That is enough rehashing of stale Miranda-bashing. Professor Weisselberg next cites studies to suggest that a concern for false confessions ought to reach beyond vulnerable suspects (especially juvenile, mentally retarded, and mentally ill ones) to certain tactics. I am open to persuasion based on evidence that particular tactics pose an unacceptably high risk of eliciting false confessions. But such a picture would require fine strokes, not a broad brush. I seriously doubt that one can generalize that every variety of deception risks inducing average adults to confess falsely. If, for example, police falsely assert that they have recovered DNA from the scene of a crime and are going to test it, such assertions will make guilty suspects fear conviction but will reinforce innocent suspects’ persistence in waiting to clear their names. If, on the other hand, police falsely tell a rape suspect that a DNA test of the rape kit has positively identified him as the rapist, the chance of a false confession would seem to be markedly higher. Those kinds of questions require more empirical study.

The real basis for Professor Weisselberg’s objection to deception seems to be a judicial and police integrity rationale, about which I’m ambivalent. There is certainly some appeal to cultivating a reputation for honesty and trustworthiness. But some kinds of deception are so trivial as to make one chuckle instead of censuring the police. David Simon famously recounted how Baltimore police pretended that a photocopier loaded with the words “Truth,” “Truth,” and “Lie” was in fact a lie detector machine. David Simon, Homicide: A Year on the Killing Streets (1991), at 49, 61 in The Miranda Debate (Richard A. Leo & George C. Thomas III eds., 1998). Such deceptive pranks hardly threaten the innocent and hardly merit the courts’ attention. Taking advantage of criminals’ stupidity and remorse does not violate the Constitution.

Likewise, Professor Weisselberg may be correct that some minimization techniques are so strong, so corrupting that they warp the confessions that they elicit. That hypothesis is plausible, but I am not sure it is proven. If it is true, it probably applies only to a subset of minimization techniques, ones that effectively put a whole story into the suspect’s mouth. There is a big difference between saying “I’m sure you’re not a bad guy and didn’t mean it,” and feeding a suspect a detailed, self-serving story through a sequence of leading questions.

IV. Encouraging Reforms

I am disappointed that Professor Weisselberg did not begin to explore possible ways to implement interrogation reforms that do not depend upon judges’ willingness to squint at the Constitution and engage in armchair empiricism. As I suggested, and he seems to agree, Miranda’s faith in the power of judicial constitutional rulings has proven unfounded. In closing, I’d like to offer some more thoughts on how nonjudicial actors might prompt and develop reforms.

First, Miranda bred an unhealthy antagonism between courts and law enforcement. Prosecutors and police want to get the guilty and free the innocent (even though, in the heat of the chase, they sometimes develop tunnel vision about who might or might not be guilty). If interrogation reforms help them to do that, they will learn to embrace them. That is why my Rebuttal noted that, despite initial resistance, police learn to embrace videotaping once they see how it captures more incriminating information and protects police against baseless charges of abuse.

Perhaps innocence commissions can prompt police departments to engage in randomized, controlled studies of various interrogation reforms, of the sort typically used in medical research. They could, for instance, learn to love double-blind interrogations, in which the interrogator does not know and so cannot disclose the key details of the crime. It is quite possible that police departments could substitute other methods for false-evidence ploys without losing truthful confessions. Or experiments might show how particular kinds of false-evidence claims create high rates of false confessions. Either finding might persuade leading interrogation trainers to reform their interrogation manuals, producing a culture change over time. Similar studies might establish presumptive cut-off times for interrogations and requirements for food, sleep, and bathroom breaks. Existing voluntariness doctrine arguably requires such scrutiny already, but Miranda’s effective safe harbor has distracted attention from these problems.

It would also be extremely valuable to study the kinds of interrogation techniques that have not induced juvenile, mentally ill, or mentally retarded defendants to confess falsely. Police interrogators ought to receive special training and direction on how to spot signs of mental illness and retardation and what techniques to use and avoid in such cases. As I suggested previously, we need scalpels, not sledgehammers. Rather than obstructing all interrogation as Miranda sought to, remedies must focus on vulnerable populations, protracted interrogations, and the disclosure of crime details to taint suspects’ confessions.

Police can also be disciplined by the prosecutors for whom, in a sense, they work. When I was a federal prosecutor in the U.S. Attorney’s Office for the Southern District of New York, our office had a policy of refusing to take any cases that depended on the testimony of certain police officers who were known to have been untruthful in the past. That sanction on “testilying” sent a powerful message that police had to remain honest if they wanted us to prosecute their cases federally. Prosecutors, prodded by embarrassing exonerations such as the Central Park Jogger case, could similarly press police for interrogation reforms. If they insisted on videotapes of confessions, or refused to take cases against mentally retarded defendants absent proper safeguards, police in their jurisdiction would take note.

One can even imagine more radical reforms. Just as the colonists had justices of the peace conduct depositions of suspects, so investigating magistrates distinct from the police could conduct double-blind interrogations today almost like depositions, tempering the worst excesses of zeal to convict. E.g., Akhil Reed Amar & Renée B. Lettow, Fifth Amendment First Principles: The Self-Incrimination Clause, 93 Mich. L. Rev. 857, 908 (1995). Even if that reform is not practical as a wholesale substitute for police interrogation, it could at least be tried for the most vulnerable populations: juvenile, mentally retarded, and mentally ill defendants.

At the end of the day, it is wrong to blame Thompkins for disregarding Miranda’s hollow edifice. Police, police trainers, prosecutors, innocence commissions, and social scientists are far better placed than courts to weed out the worst interrogation methods and propagate the best. The innocence movement has created the political moment for such reforms, from which fruitless debates over Miranda should not distract us. I thank Professor Weisselberg as well as the editors for our illuminating disagreement and hope that this dialogue will draw our readers into this conversation as well.